Project Summary Emotional dysfunction is at the core of many psychiatric disorders, with epidemiological studies noting the growing prevalence of fear, anxiety, post-traumatic, and mood disorders in the United States. The overall goal of the proposed project is to characterize the neural mechanisms that underlie dysfunctional aversive learning in anxiety disorder patients. The failure to discriminate between threatening and safe contexts is at the core of many psychiatric problems, with overgeneralization proposed as a defining feature of anxiety disorders. In this project, we (1) translate a robust basic science paradigm testing different modes of aversive learning into the clinical arena, and (2) establish a quantitative approach for measuring inter- individual differences in aversive learning. The approach relies on a robust index of aversive learning -- the steady state visual evoked potential (ssVEP) -- which is measured by presenting a visual cue at a specific frequency (e.g., 15 times per second, 15 Hz) that then elicits neural activity at the same driving frequency. When a cue is reliably followed by an aversive event (i.e. CS+), the amplitude of the ssVEP is heightened, compared to the amplitude of the frequency at which a safe cue (i.e. CS-) is presented, providing an exquisite measure of fear learning. The current project assesses generalization learning in anxiety disorder patients, by presenting an aversively conditioned CS+ together with safe cues that vary in similarity to the CS+. Pilot data with healthy participants shows neural sharpening (i.e., good discrimination of CS+) in sensors placed over occipital cortex, and neural generalization (enhanced responses to safe cues most similar to the CS+) over parietal cortex, and the proposed study measures aversive learning in 100 participants (80 anxiety disorder patients and 20 healthy controls). Using the ssVEP (as well as startle and self-reports), aversive sharpening and generalization are computed for each participant and for each dependent measure using the norm (Euclidian distance) of the difference between weights modeling sharpening or generalization functions and each participant's z-transformed means across safe cues that vary in similarity to the CS+, producing a single quantitative index indicating the degree of generalization and sharpening for each measure. The project explores the overarching hypothesis that inter-individual differences in the generalization of aversive learning- - quantitatively elucidated using the ssVEP-- are dimensionally related to key features of psychopathology across diagnostic groups. Taken together, this project will provide key information regarding dysfunction in aversive learning, which plays a role in the etiology and maintenance of anxiety, as well as form the basis for developing a novel technique useful in clinical assessment and post-treatment contexts.